Social media firms step up age checks

Barbara Ortutay, MDT/AP Technology Writer

Of the child safety measures agreed to in Meta’s landmark legal settlement, no category contains more detailed requirements than its commitments on age assurance — and for good reason.

No matter how well the protections work, they are of little use if Meta cannot tell which users are kids.

Determining which accounts belong to children is also the linchpin of measures other companies are ramping up amid growing public and regulatory pressure.

The Meta settlement — with protections that could become industry standards — gives the company a year to refine its approach.

But options are limited, and better age-verification technology presents steep challenges, including privacy implications for children and adults who may have to prove they are over 18.

The $18 billion settlement resolves state claims that Meta harmed children’s mental health by deliberately designing Instagram and Facebook to hook young people. The agreement involves 48 states, the District of Columbia and U.S. territories.

Offline, people use government-issued IDs to prove they are over 18 or 21. While this works online, many are uncomfortable giving Meta or Google access to a driver’s license or passport.

Meta accepts IDs to check ages and promises to delete them within 30 days. Beyond IDs, however, a growing number of companies use artificial intelligence to determine users’ ages. This is not true verification. AI can estimate age, often accurately, but cannot verify a birth date.

Google uses an age-assurance system for YouTube that relies on AI to distinguish adults from minors based on watch histories. Meta does the same.

On Instagram, “contextual clues” include posts about birthdays or school, which posts users like, comment on or linger over, who their friends are and when they log in. Someone offline during school hours, for instance, is likely school-aged.

Meta says its AI also examines photos for visual cues, including “height or bone structure,” to estimate general age without identifying the person.

Under the settlement, Meta agreed to strengthen its technology using its own and third-party tools, with audits. The agreement sets goals for false positives — minors identified as over 18. If a user is identified as under 13 and removed, Meta will check the ages of that user’s friends.

Identity-verification companies such as Yoti and Persona use video selfies to estimate age. Roblox requires players to take a video selfie if they want to chat with others and says the videos are deleted after processing.

Some users report inaccuracies, while experts question the reliability of facial age estimation, especially for women and certain racial or ethnic groups. Roblox says that for people ages 5 to 25, its system can estimate age within one or two years.

Age checks aim to keep children out of unsafe online spaces. But critics warn they could create a less secure, private and free internet, restricting access not only to pornography but also news, health information and anonymous speech.

What is harmful to minors is subjective, raising First Amendment concerns. People could eventually be required to verify their ages to access anything from Netflix to a neighborhood blog.

The Electronic Frontier Foundation warns that every system requires sensitive personal information linking offline identity to online activity. Once collected, that data can be leaked, hacked or misused. Protecting children matters. Building an identity checkpoint across the internet carries risks.

[Abridged]